Skip to Main content Skip to Navigation

Méthodologies de conception optimale de systèmes de conversion électromécanique

Abstract : This thesis proposes contributions in design methodologies for electromechanical systems, in particularly electrical machines. The proposed strategy allies precision and reduction in simulation time. Thus, two themes were studied. The first theme concerns electromagnetic modeling by means of semi-numerical models. Reluctance’s network topology is proposed in order to analyze performances of permanent magnet axial flux machines. It can take into account the temporal aspect and the non-linearity of ferromagnetic materials. Finite element models were established in order to compare and validate the proposed modeling strategy. Thus, the small error between models (less than 5%) shows the efficiency of the models. Sizing based on optimization techniques, offering a compromise between final solution and computation time, are treated in the second theme of this thesis. Space Mapping is known as an efficient optimization technique, which enables exploiting costly models without being prohibited by computation time. Two derived techniques are proposed and applied to design the axial flux machines. The first, proposes the modeling of the error between models with different precision through radial basis function. The second embeds error correction in the magnetic equivalent circuit resolution in order to couple linear and non-linear models of an electrical machine.
Complete list of metadatas

Cited literature [211 references]  Display  Hide  Download
Contributor : Maya Hage-Hassan <>
Submitted on : Wednesday, October 2, 2019 - 1:46:41 PM
Last modification on : Thursday, September 17, 2020 - 12:26:02 PM


Files produced by the author(s)


  • HAL Id : tel-02303552, version 1


Maya Hage-Hassan. Méthodologies de conception optimale de systèmes de conversion électromécanique. Energie électrique. Université Paris Sud (Paris 11), 2014. Français. ⟨tel-02303552⟩



Record views


Files downloads